The text-to-video AI boom has really taken off in recent months, with the only downside being that there's nothing like OpenAI Sora to try out yet. If you can't wait, a new competitor called Dream Machine has just arrived, and you can try it out now.
Necessities

Video Resolution: 1360×752
Clip length: 5 seconds
Training data: Unknown
Fee: Free plan offers 30 generations per month, paid plans start at $29.99 per month
Dream Machine is developed by Luma AI, which previously released an app that let you take 3D photos on your iPhone, and now the company is focusing on generative video, and it's free to use right now with a Google account, with a few caveats.
The main issue is that Dream Machine seems to be a bit overloaded at the time of writing. Currently, the site displays a banner saying “Generation will take 120 seconds” and “Due to high demand, your request will be queued.” The text prompt took over 20 minutes to process, but the results (below) are very impressive.
Dream Machine's output is limited in length and resolution compared to OpenAI's Sora or Kling AI, but it's a good test to see how these services work. The clips Dream Machine generates are 5 seconds long and have a resolution of 1360 x 752. Simply enter the prompt into the search bar and wait for it to appear in your account, after which you can download the watermarked version.
There was a long wait for results (which should improve once the initial demand subsides), but we ended up with a clip that was pretty close to what we had in mind for our theme of “close-up of a dog wearing sunglasses driving a car in Las Vegas at night.”
Dream Machine's free plan is limited to 30 generations per month, but if you need more there's Standard (120 generations, $29.99 per month, about £24, AU$45), Pro (400 generations, $99.99 per month, about £80, AU$150) and Premier (2,000 generations, $499.99 per month, about £390, AU$750).
A preview of upcoming AI videos
As with most generative AI video tools, questions remain about what data Luma AI was trained on, which means its potential is likely limited outside of personal use and improving your GIF game. It's also not the first free text-to-video tool we've seen; Runway's Gen 2 model was released out of beta last year.
Dream Machine's website also notes that the tool has technical limitations when it comes to handling text and motion, so there will be plenty of trial and error involved, but as a trial for upcoming, more advanced (and no doubt more expensive) AI video generators, it's certainly a fun tool to try out.
This is especially true given that other alternatives such as Google Veo currently have long waiting lists, while more powerful models such as OpenAI's Sora (which can generate 60-second videos) won't be available until later this year, while Kling AI is currently only available in China.
This will certainly change once text-to-video generation becomes mainstream, but until then, Dream Machine is a good place to practice (if you don't mind waiting a while for results).
